T06P07 - China’s Subnational Government Relation and Policy Implementation: In the Shadow of Central-local Paradigm

Topic : Policy Implementation

Panel Chair : Ciqi Mei - cmei@tsinghua.edu.cn

Objectives and Scientific Relevance of the panel

Call for papers

Session 1 Policy and Politics in Subnational China: Trust, Loyalty and Awareness


Ciqi Mei - cmei@tsinghua.edu.cn - Tsinghua University - China

Jiangnan Zhu - zhujn@hku.hk - The University of Hong Kong - Hong Kong, (China)

Do anticorruption campaigns boost government trust in China?

siqin kang - skang@connect.ust.hk - Hong Kong University of Science and Technology - Hong Kong, (China)

Jiangnan Zhu - zhujn@hku.hk - The University of Hong Kong - Hong Kong, (China)

Voluminous studies have established that corruption will undermine political trust and states that succeed in reducing corruption are more likely to be valued by citizens (Diamond, 1999). In the past decades, anti-corruption campaign became a most prominent one in fighting with corruption. However, research examining the campaign’s impact on government trust has been limited and findings are inconsistent.

The unprecedented large scale anticorruption campaign launched by Chinese government in the end of 2012 provided a good opportunity for studies. In the past years, more than 100 officials at/above vice minister/provincial-level and more than 10,000 officials at/above department/bureau-level officials have been charged for corruption. We attempt to explore the relationship between the anti-corruption campaign and political trust providing a new perspective and unique data.

Existing literature holds that policies influence agencies generally in two ways, imposing real changes or transmitting powerful signals to the agencies (Carpenter, 1996). Therefore, we argue that anticorruption campaigns should boost public trust, if they can, mainly through two effects: signaling government’s commitment to anticorruption (i.e. “signaling effect”) and actual reduction or deterrence of corruption (i.e. “actual effect”). It is important to distinguish the two effects, because the actual reduction of corruption is more desirable and the signaling effect is more like quenching a thirst with poison. We focus on public trust to local government, because local officials are accessible to ordinary people, and thus both the signaling effect and actual effect may work in the process. A Difference-in-Differences research design with two wave of survey data (2012 and 2014) from China Family Panel Studies (CFPS) is used. We collected number of arrested officials in each municipality during the survey period to measure the strength of anticorruption in each locality.

For the actual effect, it is expected that those who have more opportunities to contact government officials are close to the site of corruption and more sensitive to the actual changing levels of corruption. We, therefore, test the difference between five sets of groups of respondents: communist party member vs. non-party member, government/SOE employee vs. non-gov/SOE employee, company managers vs. ordinary employee, urban employee vs. farmers, and cadre vs. non-cadres. We find that none of the groups having more contacts with government officials show more trust to government when more local officials are arrested in a locality. In other words, the actual effect is not a major channel to increase public trust in government.   

For the signaling effect, because its foundation is government credibility (Licari and Meier 2000), we expect that people who have access to more information other than official one are more likely to be critical to government and less affected by the signaling effect of anticorruption campaigns. Thus, we test the differences of four sets of groups: people with different levels of education, age group, Internet users, and levels of trust before campaign. We find that those who have better education or more access to alternative information sources are less responsive to the number of arrested local officials. Thus, signaling effect hypothesis is supported.

Fiscal Decentralization and Political Trust in China

Shengqiao LIN - linsq15@mails.tsinghua.edu.cn - School of Public Policy and Management, Tsinghua University - China

Political trust is an essential basement of legitimacy of a modern state. Past decades in the developed industrialized countries, political trust stayed stable in the government keep declining. Decentralization, with its promises to improve the relation between the government and its citizenry and increase the efficiency of government service, is considered to be one of such reform policies.


Recent studies focus on the association between fiscal decentralization and political trust. Some attach importance to the impact that decentralization improves relations between the government and its citizenry. However, some concentrate more on indirect influence that decentralization helps to promote the performance of government and impact political trust. These theories base on the assumption that fiscal decentralization synchronizes with administrative and political decentralization. The asymmetric decentralization in China makes the effect of fiscal decentralization on political trust more complicated.


This research try to combine these two approaches to reveal the impact of fiscal decentralization on political trust in China. One the one way, decentralization contributes to the improvement of the relation between government and its citizenry. Three potential mechanisms might exist. Firstly, based on classical decentralization theory, local governments with more fiscal autonomy are supposed to satisfy people more effectively with a shorter distance of information. In the meanwhile, cities with more degree of fiscal decentralization are expected to attract qualified personnel and strengthen their capacity. Lastly, in cities with more autonomy in decision making, citizen would have more potential chances in political participation.


One the other hand, with more autonomy they have in decision making, they would use their fiscal capacity in a way more aligned with their preference. Local governments show different preferences on economic development and public service. As asymmetric decentralization in China, central governments keep major personal administration and evaluate government officials’ performance on their economic development mainly. With more autonomy, regional governments would spend more on developments rather than public service, which is so-called expenditure distortion. However, the proportions of different fiscal expenditure have diverse influence on political trust. After continued economic development in China last decades, people value government’s effort on public service more than economic development. The mismatch between people’s evaluation and government’s preference leads to continued declining political trust in local governments in China.


This paper discloses the casualty on how fiscal decentralization influences political trust. First, it combined individual-level from China General Social Survey (2010) and city-level fiscal data. Quantitative analysis with multilevel ordinal regression model serves to show the effect of fiscal decentralization on political trust. Secondly, this paper uncovered the causal mechanism by using different fiscal expenditure preference as mediating variables. As symmetric decentralization in China, governments’ fiscal preference goes contrary to people’s evaluation. The direct and indirect effects of fiscal decentralization are supposed to be opposite in China.


This study contributes a structural view to explain the downturn of political trust to local governments in China. Taking fiscal decentralization system in China as a case, the preference-evaluation model serves to understand the relation between fiscal decentralization and political trust especially in governments of different levels.

Duck’s paddling: the sub-provincial policy distortion in the Case of China’s ESER Policy Implementation

Ciqi Mei - cmei@tsinghua.edu.cn - Tsinghua University - China

Shaowei Chen - chen-sw12@mails.tsinghua.edu.cn - 21st China Center, UC San Diego - United States

The China scholarship on policy implementation has long noted policy distortion happened at various local levels. (O’Brien & Li, 1999; Kostka, 2014, 2016; Ran, 2013, 2015; Gao 2015). However, previous studies have mostly treated indifferently the distortion happened at provincial level and those at sub-provincial levels. Such treatment could be problematic as provincial governments and sub-provincial government are facing different implementation incentives and supervisions. For the provincial governments, supervisions on them come directly from the central government who has a clear policy commitment; for the sub-provincial governments, however, supervisions come from the provincial governments whose primary target is to have the sub-goals, -assigned to them from the center-, met. A conjecture is therefore that distortion at the sub-provincial level and might take a different form than that taken by a provincial government who is facing strict supervision from the center. As a result, while provincial government might well report satisfactory policy implementation, actual distortion to the policy commitment could happen at the sub-provincial level, like a duck floating poised on surface but paddling like hell beneath the water.


Implementation of “energy saving and emission reduction” (ESER) from 2006 might evidence such duck metaphor. Based upon the official report, most provinces with few exceptions in China have successfully met five-year goal of center’s mandate on ESER program. However, the mission to ESER program, i.e. to build a “greener” society, still falls short as felt by ordinary citizens. Despite the pretty grade sheet, distortion might have happened at the subnational level. This paper aims to identify distortion mechanism at the sub-provincial level. Specifically, while the provinces mostly report their ESER targets have been met, this paper finds out these numbers are aggregated output of policy behaviors distorted from the center’s policy mission at prefectural level.


Specifically, we look at prefectural government’s policy attention devoted to ESER as our dependent variable. We develop a measurement of policy attention by coding the annual intensity of ESER-related news reports in the official newspaper owned by prefectural governments. We find out that cities with higher economic development ranking would allocate less attention on ESER policy and vice versa. In other words, those cities who need put more efforts into ESER programs might defy the center’s mandate safely as others could nevertheless take over their share. In the end, the provincial government could meet their designated goal but policy mission set by the central government falls shorts. Various econometric models confirm robustness of our findings.

Bureaucratic Discretion and Behavioral Logics of Intermediate Governments

Xiao Shiyang - leaforlife@sina.com - School of Public Policy and Management, Tsinghua University - China

By borrowing ideas from information economics, the structure of most government organizations can be simplified as a three-tier principal/supervisor/agent model. Due to the high cost for central government (principal) to supervise performance at the bottom level, the intermediate governments (i.e.: provincial governments in our research) as “supervisors” are entitled with large discretion to decide what and how to implement a top-down central policy. Researches trying to explain the variation in the use of such discretion emphasize a close link between environmental factors and bureaucratic behaviors. However, few studies show how these environmental characteristics interact with each other and co-influence the behavioral logics of government agencies.


In our research, we emphasize two types of environmental factors: policy environment and political environment. The former refers to policy attributes; The latter refers to the network of authority relationships between different agencies and groups. Our study try to show how policy & political environments impact the behaviors of intermediate governments. More importantly, we try to show how these different factors interact with each other and co-influence bureaucratic activities.


The "policy environment" in our research includes two factors: “policy impact on the core interests of central government (principal)” and “policy burden on local governments (agents)”. The political environment here refers to “provincial reliance on central government” (measured by financial freedom). To explore the influence of these environmental factors and their interaction on government behaviors, we focus on central social regulatory policies from 2003-2012 in China, and to see whether and how fast a province responds to central government by releasing a corresponding document. We will show how the above environmental factors and their interaction impact provincial responses to central policies.


Our research tries to demonstrate that intermediate governments face dual behavioral logics: pleasing principal & protecting agents. In other words, provincial governments will be more willing to respond if a policy has an impact on the core interests of central government, and will be reluctant to respond if a policy imposes burden on local agencies. The above findings justify our  hypotheses about the impacts of policy environment on bureaucratic behaviors. Furthermore, we will also show how the above behavioral logics are influenced by the extent to which a certain province relies on central government (political environment). We try to prove that provinces which rely heavily on central government will be more reluctant to protect agents, while be more willing to release “symbolic documents” in order to please principal. 

Beyond bureaucracy: The Effects of Environmental Policy Experimentation on Citizens’ Environmental Awareness in Urban China

Yue Guo - jason.guoyue@gmail.com - Harvard University - United States

JIE WANG - 2296929197@qq.com - China

Zhilin Liu - zhilinliu@mail.tsinghua.edu.cn - Tsinghua University - China

A large volume of literature has studied the role of policy experimentation in China’s policy change and governance transformation. Typically from the perspective of the central-local relationship, previous studies mainly focused on the intended effects of policy experimentation, such as on policy formulation at central level (Heilmann, 2008; Mei and Liu, 2014) and policy diffusion among local governments (Zhu and Zhang, 2015). To our best knowledge, however, little research has been conducted on the “unintended” effects of policy experimentation beyond bureaucracy, such as the effects of policy experimentation on how the public perceive a policy issue and even change their behavior toward the issue. This policy feedback mechanism has drawn increasing attention among public policy scholars in the past decade, who are intrigued by the “unintended effect” of policies on governance. This concept of policy feedback has been also gradually adopted as a new approach to study policy process and explain how existing policies shape public opinion--treating policies as causes rather than merely effects.


In this study, we conduct a multi-level analysis to empirically investigate the extent to which different policy experimentation models – i.e. top-down and bottom-up – may have different impacts on public opinion. We use environmental governance as a case and identify 13 sustainability-related experimental programs initiated by various central government ministries. After extensive review of policy documents, we further classify all experimental programs into pilot-city programs (shidian chengshi) and model-city programs (mofan chengshi), each representing the top-down and bottom-up models for policy experimentation in China. The model-city program typically is initiated by the central government to award and promote local performance in sustainability whereas the pilot-city program typically aims at advocating certain policy goals or testing specific policy measures championed by the central government. We build a database that combines the 2010 China General Social Survey (CGSS) with a city-level database capturing how many titles one city was been honored to conduct pilot-city programs or model-city programs related to environmental sustainability by central government ministries during 2006-2010? Included control variables would be personal attributes– e.g. age, gender, education, household income and environmental knowledge, and city-level contextual factors including economic development and environmental quality.


Our empirical results suggest that two opposite environmental experimentation models - top-down and bottom-up - affect public environmental opinion in different ways. Urban residents tend to show higher level of environmental awareness to environmental issues in cities that have been pioneers as central government designated model cities in environmental policy while they are more actively engaging in environment-friendly behavior in cities with more pilot titles. Our findings show that the policy feedback effect brought by bottom-up approach is more effective and positive on public opinion due to a mobilization mechanism. Public participation is more encouraged by the bottom-up environmental policy experimentation approach, which could promote environmental awareness to environmental issues. When public participate environmental governance, they are shaped by government policies as well.

Export PDF